Home/Technology & AI/AI and space travel are hype that won't deliver
Technology & AIDidn't happen

ChatGPT-style AI is not any good and will not get any better; AI hype is a scam.

Said Apr 1, 2025 in Civilization #43: The Structure of Scientific Revolutions @ 51:53

Timeline Resolved 4 months after he said it

Timeframe unspecifiedSaid again 10×

Verdict Didn't happen

ChatGPT-style models kept improving after April 2025: OpenAI released GPT-5 on 2025-08-07 with large benchmark gains, and OpenAI followed it with GPT-5.1, 5.2, 5.4, 5.5 and later GPT-6 releases through 2026.

en.wikipedia.orgopenai.com

What he said

So, I'm sure you've all used ChatGPT. Is it any good? No, it's not. Will it get any better? No, it won't, okay?
Read it in the transcript, in context →

He keeps saying this

Restated in 9 videos as AI and space travel are hype that won't deliver

More from this video

Technology & AI

Artificial general intelligence, nanotechnology and genetic immortality/designer babies are not achievable within current science; they are illusions or scams.

As of Oct 2026 none of AGI, mature molecular nanotechnology or genetic immortality has been achieved, but AI capabilities continue to advance; the claim that they are 'illusions' cannot be settled yet.

AI and space travel are hype that won't deliverCivilization #43: The Structure of Scientific Revolutions @ 51:25
Open
Said Apr 1, 2025no deadline
Predictive History

An independent scorecard of every forward-looking claim Professor Jiang Xueqin made on the @PredictiveHistory YouTube channel. Not affiliated with the channel.

Updated 2026-10-16 · 757 calls from 180 transcripts.

About · Method · Contact · Videos · Data

How calls are found and checked

Each transcript was read in full by an AI model to pull out predictions, with the exact quote and timestamp. Each one was then checked with web search against events up to the update date, with sources linked.

How the hit rate works

Spot a mistake? Report it. Hit rate = (came true + ½ partly) ÷ resolved calls. Claims that were already underway when he made them are left out. Open calls haven't reached a verdict; Unresolvable ones are too vague ever to check. Verdicts are AI-assisted and may contain errors.